Abstract:

In order to verify the accuracy of the graphics measured by a non-contact scanner, the badk-curves of adult males in the group of 170/88A were studied, then the function fitting and the accuracy verification was performed based on the manual measurement, the circumscribed circle along its curve was drawn. The back-curve were fitted by a polynomial function, and the R-squared of cubic polynomial function is 0.995 62. Considering the concision, the cubic polynomial function is determined to be the standard reference curve.  The correlation verification shows that the back-curve length, back length and chest depth are related, the fitting function is converted into a parameter function, and based on the functional relationship among them, zoomed was performed according to the proportional relationship, and further the back-curve length was calculated from the back length and chest depth of a sample in the same group, and compared with the length of the fitted curve. This experiment results show that the error of the back-curve function fitted is smaller than 1 cm, which is far smaller than the class difference for clothes, and the accuracy is enough to satisfy the requirement to the non-contact human body measurement for clothes.

Key words: non-contact human body measurement, back curve, cubic polynomial fitting, correlation verification